The git diff (specifically "git diff master claude", with and without --stat) and perhaps the aforementioned CLAUDE_PROGRESS.md file (Claude written) are going to be your best bet at this point.
It's still an audacious experiment. Deterministic simulation testing is such an incredibly valuable bug finder/fixer that its worth trying to resurrect Gosim from being abandoned, even if its plumbing into the Go runtime is Byzantine and hard to grok -- at least for me. On Thursday, February 26, 2026 at 3:08:27 PM UTC-3 Mike Crnkovich wrote: I would recommend better care with the commit messages. I went to take a look at the changes that were being added, and it was tough to follow. On Thu, Feb 26, 2026, at 10:26 AM, Jason E. Aten wrote: Heh. While it lots of situations I agree whole heartedly, fortunately in this particular situation with Gosim we have a readily available correctness Oracle: the normal (non-deterministic) Go runtime execution can be readily compared. If one of the many inconsistent executions from the normal runtime matches the consistent, reproducible execution from Gosim, then we know the linkname translation minor repairs worked. On Thursday, February 26, 2026 at 9:39:02 AM UTC-3 Wojciech S. Czarnecki wrote: Dnia Wed, 25 Feb 2026 16:55:45 -0800 (PST) "Jason E. Aten" napisał/a: > Claude-code fix > [...] > I make no guarantees that it is correct, > [...] it > does appear to work I eagerly am awaiting similar text to appear on pre-surgery "informed consent" waiver papers. Like "I am aware my chest will be opened by superb surgical robot maintained by Clunky-code. We have seen it operating and it does appear to work". -- Wojciech S. Czarnecki << ^oo^ >> OHIR-RIPE -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion visit https://groups.google.com/d/msgid/golang-nuts/7668250b-5d4c-4ba2-b379-98bda1136f61n%40googlegroups.com.
